[1] "Running analysis on CZs (cz_id)."
Regression Results
Introduction
The following document summarises the progress made thus far on Chapter 1: Local Fiscal Risks of Decarbonisation of my DPhil. The work aims to pursue a better understanding of how industrial transformation impacts local well-being. From an original interest in looking at all aspects of local public finance, the project has narrowed to focus on expenditure on public education and its connection to industrial prosperity and transformation.
Current strategy/research plan: 1. Outcome: Educational Expenditure 2. Treatment (endogenous): Wages, Economic Growth, Property Values, Property Taxes 3. Instrument: Industry Shares of Employment in high vs. low wage growth industries/sectors - plausible exogeneity comes from industrial shares. Need to justify the choice of base-year for the shift-share instrument such that industries were “present” but still nascent (this is important because there are likely to be certain industries that “cropped up” post-baseline completely unrelated to the industrial composition before right?)
Recall the work from the previous meeting:
- After reading more work on US economic geography, it became clear that aggregating counties up to commuting zones was the better choice for analysis at sub-state level as these areas more accurately represent local labour markets/economies/commuting zones (Fowler et al. 2024, David Dorn’s Resource Page).
- Below, I provide some baseline regressions to demonstrate the relationships between key variables in the dataset.
- Next, I turn to an instrumental variable application in which I use coal mine counts and production volumes as an instrument for property taxes. Coal mine counts do not serve as good instruments but coal production passes relevance and exogeneity restrictions. I believe a strong argument can be made for the exclusion restriction to be satisfied. I provide supporting statistical tests for all that demonstrate the unfitness of coal mine counts but fitness of coal production. Along this line, we can hopefully discuss other sources of variation in industrial/economic productivity that might lead to property value spirals (positive or negative) to test the property tax channel.
- I identify declining vs. growing regions by estimating commuting-zone growth rates conditional on state and national level growth rates. Using this distinction (on both a per capita and total gdp bases and a lenient vs. stringent magnitude threshold), I rerun the key regressions identified in steps 2 and 3 on the subgroups (declining and growing regions).
Note: Any warnings about “missing observations” or “NA being removed” relates to the lags incorporated, except in the Bartik estimations.
Data
All data used is reported annually at the commuting zone level. Therefore, no time-invariant variables are included (apart from the State in which a commuting zone is in, which is made time-variant through the inclusion of a state-level trend in various models). 636 commuting zones in 40 states between 2001-2021.
Expenditure and Revenue: The dependent variables of interest come from Willamette University’s Government Finance Database. The data includes commuting-zone level revenue and expenditure on public education including disaggregated values by revenue source (federal, state, or other intergovernmental revenue) and expenditure item (lunches, wages, debt). All values are reported in real US dollars. The data for property taxes collected used in regressions below also come from this dataset. Expenditure on vocational training and from Educational Service Agencies (ESAs) are also sourced from this dataset.
GDP Controls: US Bureau of Economic Analysis. Values are also reported in current US dollars (real GDP values exist). The controls used in the below are total, private industry, and oil, gas, mining & quarrying commuting zone-level GDP.
Population controls: US Census Bureau.
Coal mine activity and production levels: Mine Safety and Health Administration
Summary statistics
All dollar values are reported in real 2017-chained thousands.
% Table created by stargazer v.5.2.3 by Marek Hlavac, Social Policy Institute. E-mail: marek.hlavac at gmail.com % Date and time: Tue, Jun 03, 2025 - 10:03:51Baseline TWFE Model
Incorporating time lags
Education expenditure has a highly relevant time dependence. The effect of increases in GDP two years prior has the greatest effect on current education expenditure, implying a delayed effect of commuting zone-level economic growth on public education expenditure. First 6 do not include state time trends; second 6 do.
\begingroup
\centering
\begin{tabular}{lccccccc}
\tabularnewline \midrule \midrule
Dependent Variable: & \multicolumn{7}{c}{(log) Elem.Ed.Exp.pp}\\
Model: & (1) & (2) & (3) & (4) & (5) & (6) & (7)\\
\midrule
\emph{Variables}\\
(log) Real GDP Priv. Industry pc & 0.0094 & -0.0144 & -0.0095 & 0.0171 & & & \\
& (0.0203) & (0.0204) & (0.0199) & (0.0198) & & & \\
(log,l1) Real GDP Priv. Industry pc & 0.0612$^{***}$ & 0.0400$^{***}$ & 0.0241 & 0.0344$^{**}$ & & & \\
& (0.0145) & (0.0153) & (0.0178) & (0.0169) & & & \\
(log,l2) Real GDP Priv. Industry pc & 0.1325$^{***}$ & 0.1073$^{***}$ & 0.0944$^{***}$ & 0.1092$^{***}$ & & & \\
& (0.0225) & (0.0232) & (0.0234) & (0.0197) & & & \\
(log) Annual Avg. Wkly. Wage & & 0.1998$^{***}$ & 0.1319$^{*}$ & 0.0483 & 0.2645$^{***}$ & 0.1592$^{**}$ & 0.1283$^{*}$\\
& & (0.0668) & (0.0699) & (0.0732) & (0.0632) & (0.0655) & (0.0679)\\
(log, l1) Annual Avg. Wkly. Wage & & 0.1779$^{***}$ & 0.1339$^{**}$ & 0.1417$^{**}$ & 0.2201$^{***}$ & 0.1621$^{***}$ & 0.1619$^{***}$\\
& & (0.0525) & (0.0578) & (0.0571) & (0.0484) & (0.0549) & (0.0540)\\
(log, l2) Annual Avg. Wkly. Wage & & 0.0215 & 0.0090 & 0.0079 & 0.1945$^{***}$ & 0.1674$^{***}$ & 0.1963$^{***}$\\
& & (0.0637) & (0.0647) & (0.0637) & (0.0678) & (0.0541) & (0.0530)\\
(log) House Price Index & & & 0.0186 & -0.0310 & & 0.0386 & 0.0003\\
& & & (0.0299) & (0.0271) & & (0.0273) & (0.0245)\\
l1\_log\_hpi & & & 0.0726$^{**}$ & 0.0573$^{*}$ & & 0.0657$^{**}$ & 0.0532$^{**}$\\
& & & (0.0327) & (0.0315) & & (0.0272) & (0.0259)\\
l2\_log\_hpi & & & 0.0776$^{***}$ & 0.0498$^{**}$ & & 0.0614$^{***}$ & 0.0306\\
& & & (0.0259) & (0.0245) & & (0.0228) & (0.0211)\\
l3\_log\_hpi & & & 0.0158 & 0.0235 & & 0.0295 & 0.0338\\
& & & (0.0237) & (0.0231) & & (0.0220) & (0.0209)\\
l4\_log\_hpi & & & -0.0118 & 0.0109 & & -0.0051 & 0.0185\\
& & & (0.0260) & (0.0230) & & (0.0221) & (0.0195)\\
l5\_log\_hpi & & & -0.0799$^{***}$ & -0.0588$^{***}$ & & -0.0897$^{***}$ & -0.0669$^{***}$\\
& & & (0.0243) & (0.0207) & & (0.0219) & (0.0190)\\
(log) State IG Rev pp & & & & 0.3600$^{***}$ & & & 0.3480$^{***}$\\
& & & & (0.0334) & & & (0.0330)\\
(log) Fed IG Rev. pp & & & & 0.0056$^{**}$ & & & 0.0055$^{**}$\\
& & & & (0.0025) & & & (0.0022)\\
\midrule
\emph{Fixed-effects}\\
unit & Yes & Yes & Yes & Yes & Yes & Yes & Yes\\
year & Yes & Yes & Yes & Yes & Yes & Yes & Yes\\
\midrule
\emph{Fit statistics}\\
Observations & 12,084 & 12,084 & 11,420 & 11,420 & 13,356 & 12,536 & 12,536\\
R$^2$ & 0.82088 & 0.82430 & 0.83203 & 0.86309 & 0.81776 & 0.82710 & 0.85772\\
Within R$^2$ & 0.07806 & 0.09567 & 0.10150 & 0.26763 & 0.08329 & 0.09465 & 0.25497\\
\midrule \midrule
\multicolumn{8}{l}{\emph{Clustered (unit) standard-errors in parentheses}}\\
\multicolumn{8}{l}{\emph{Signif. Codes: ***: 0.01, **: 0.05, *: 0.1}}\\
\end{tabular}
\par\endgroup
\begingroup
\centering
\begin{tabular}{lccccc}
\tabularnewline \midrule \midrule
Dependent Variable: & \multicolumn{5}{c}{(log) Elem.Ed.Exp.pp}\\
Model: & (1) & (2) & (3) & (4) & (5)\\
\midrule
\emph{Variables}\\
state\_share & 0.3807 & 0.4409 & -0.4090 & 0.1497 & -0.6038\\
& (0.6027) & (0.6279) & (0.6709) & (0.4649) & (0.4585)\\
(log) Real GDP Priv. Industry pc & -0.2444$^{***}$ & -0.2447$^{***}$ & -0.1677$^{**}$ & & \\
& (0.0775) & (0.0814) & (0.0778) & & \\
(log,l1) Real GDP Priv. Industry pc & 0.1036$^{*}$ & 0.0778 & -0.0283 & & \\
& (0.0614) & (0.0630) & (0.0684) & & \\
(log,l2) Real GDP Priv. Industry pc & 0.3252$^{***}$ & 0.2954$^{***}$ & 0.2816$^{***}$ & & \\
& (0.0801) & (0.0785) & (0.0724) & & \\
state\_share $\times$ (log) Real GDP Priv. Industry pc & 0.4415$^{***}$ & 0.3919$^{***}$ & 0.2482$^{**}$ & & \\
& (0.1274) & (0.1335) & (0.1258) & & \\
state\_share $\times$ (log,l1) Real GDP Priv. Industry pc & -0.1160 & -0.1056 & 0.0551 & & \\
& (0.0996) & (0.1014) & (0.1082) & & \\
state\_share $\times$ (log,l2) Real GDP Priv. Industry pc & -0.4256$^{***}$ & -0.4044$^{***}$ & -0.3903$^{***}$ & & \\
& (0.1235) & (0.1222) & (0.1031) & & \\
(log) Annual Avg. Wkly. Wage & & -0.0579 & -0.0523 & -0.0169 & -0.0695\\
& & (0.1949) & (0.1995) & (0.2152) & (0.2062)\\
(log, l1) Annual Avg. Wkly. Wage & & 0.2640$^{*}$ & 0.1269 & 0.2405 & 0.1220\\
& & (0.1551) & (0.2255) & (0.1549) & (0.2182)\\
(log, l2) Annual Avg. Wkly. Wage & & 0.1513 & 0.0969 & 0.3747$^{*}$ & 0.3149$^{*}$\\
& & (0.1657) & (0.1791) & (0.1979) & (0.1706)\\
state\_share $\times$ (log) Annual Avg. Wkly. Wage & & 0.5684$^{*}$ & 0.4082 & 0.5184 & 0.3950\\
& & (0.3118) & (0.3189) & (0.3406) & (0.3260)\\
state\_share $\times$ (log, l1) Annual Avg. Wkly. Wage & & -0.1879 & -0.0520 & -0.0884 & 0.0042\\
& & (0.2504) & (0.3474) & (0.2472) & (0.3352)\\
state\_share $\times$ (log, l2) Annual Avg. Wkly. Wage & & -0.3645 & -0.2613 & -0.5583$^{*}$ & -0.4675$^{*}$\\
& & (0.2823) & (0.2785) & (0.3234) & (0.2708)\\
(log) House Price Index & & & -0.1954 & & -0.2105$^{*}$\\
& & & (0.1300) & & (0.1219)\\
l1\_log\_hpi & & & 0.1465 & & 0.1385\\
& & & (0.1673) & & (0.1373)\\
l2\_log\_hpi & & & 0.4083$^{***}$ & & 0.3942$^{***}$\\
& & & (0.1190) & & (0.0990)\\
l3\_log\_hpi & & & -0.0575 & & -0.0163\\
& & & (0.0980) & & (0.1096)\\
l4\_log\_hpi & & & -0.1427 & & -0.1251\\
& & & (0.1095) & & (0.0925)\\
l5\_log\_hpi & & & -0.0397 & & -0.1044\\
& & & (0.0831) & & (0.0851)\\
state\_share $\times$ (log) House Price Index & & & 0.4394$^{**}$ & & 0.4690$^{**}$\\
& & & (0.1983) & & (0.1870)\\
state\_share $\times$ l1\_log\_hpi & & & -0.1329 & & -0.1294\\
& & & (0.2474) & & (0.2050)\\
state\_share $\times$ l2\_log\_hpi & & & -0.5463$^{***}$ & & -0.5355$^{***}$\\
& & & (0.1807) & & (0.1493)\\
state\_share $\times$ l3\_log\_hpi & & & 0.0967 & & 0.0493\\
& & & (0.1542) & & (0.1707)\\
state\_share $\times$ l4\_log\_hpi & & & 0.1836 & & 0.1638\\
& & & (0.1668) & & (0.1428)\\
state\_share $\times$ l5\_log\_hpi & & & -0.0457 & & 0.0454\\
& & & (0.1276) & & (0.1327)\\
\midrule
\emph{Fixed-effects}\\
unit & Yes & Yes & Yes & Yes & Yes\\
year & Yes & Yes & Yes & Yes & Yes\\
\midrule
\emph{Fit statistics}\\
Observations & 12,084 & 12,084 & 11,420 & 13,356 & 12,536\\
R$^2$ & 0.84836 & 0.85161 & 0.86210 & 0.84454 & 0.85655\\
Within R$^2$ & 0.21950 & 0.23623 & 0.26234 & 0.21800 & 0.24887\\
\midrule \midrule
\multicolumn{6}{l}{\emph{Clustered (unit) standard-errors in parentheses}}\\
\multicolumn{6}{l}{\emph{Signif. Codes: ***: 0.01, **: 0.05, *: 0.1}}\\
\end{tabular}
\par\endgroup
Incorporating state-level trends
The below take the Education Expenditure ~ GDP models and incorporate deterministic state time trends.
\begingroup
\centering
\begin{tabular}{lccccccc}
\tabularnewline \midrule \midrule
Dependent Variable: & \multicolumn{7}{c}{(log) Elem.Ed.Exp.pp}\\
Model: & (1) & (2) & (3) & (4) & (5) & (6) & (7)\\
\midrule
\emph{Variables}\\
(log) Real GDP Priv. Industry pc & 0.0152 & 0.0021 & 0.0054 & 0.0253 & & & \\
& (0.0192) & (0.0195) & (0.0203) & (0.0200) & & & \\
(log,l1) Real GDP Priv. Industry pc & 0.0550$^{***}$ & 0.0405$^{***}$ & 0.0256 & 0.0355$^{**}$ & & & \\
& (0.0150) & (0.0154) & (0.0179) & (0.0170) & & & \\
(log,l2) Real GDP Priv. Industry pc & 0.1044$^{***}$ & 0.0917$^{***}$ & 0.0815$^{***}$ & 0.1110$^{***}$ & & & \\
& (0.0249) & (0.0245) & (0.0243) & (0.0213) & & & \\
(log) Annual Avg. Wkly. Wage & & 0.0902 & -0.0291 & -0.0329 & 0.2033$^{***}$ & 0.0758 & 0.1010\\
& & (0.0652) & (0.0678) & (0.0708) & (0.0595) & (0.0626) & (0.0655)\\
(log, l1) Annual Avg. Wkly. Wage & & 0.1676$^{***}$ & 0.1200$^{**}$ & 0.1276$^{**}$ & 0.1840$^{***}$ & 0.1355$^{**}$ & 0.1370$^{**}$\\
& & (0.0518) & (0.0560) & (0.0567) & (0.0481) & (0.0543) & (0.0547)\\
(log, l2) Annual Avg. Wkly. Wage & & 0.0289 & 0.0803 & 0.0650 & 0.1607$^{**}$ & 0.2098$^{***}$ & 0.2432$^{***}$\\
& & (0.0629) & (0.0638) & (0.0593) & (0.0758) & (0.0571) & (0.0544)\\
(log) House Price Index & & & 0.0628$^{**}$ & 0.0272 & & 0.0676$^{***}$ & 0.0466$^{*}$\\
& & & (0.0280) & (0.0261) & & (0.0248) & (0.0238)\\
l1\_log\_hpi & & & 0.0522 & 0.0407 & & 0.0466$^{*}$ & 0.0385\\
& & & (0.0321) & (0.0309) & & (0.0272) & (0.0260)\\
l2\_log\_hpi & & & 0.0712$^{***}$ & 0.0409$^{*}$ & & 0.0550$^{**}$ & 0.0276\\
& & & (0.0254) & (0.0248) & & (0.0223) & (0.0211)\\
l3\_log\_hpi & & & 0.0262 & 0.0244 & & 0.0281 & 0.0273\\
& & & (0.0237) & (0.0228) & & (0.0216) & (0.0210)\\
l4\_log\_hpi & & & -0.0032 & 0.0160 & & -0.0037 & 0.0180\\
& & & (0.0260) & (0.0227) & & (0.0220) & (0.0190)\\
l5\_log\_hpi & & & -0.1303$^{***}$ & -0.0601$^{***}$ & & -0.1301$^{***}$ & -0.0627$^{***}$\\
& & & (0.0239) & (0.0212) & & (0.0217) & (0.0196)\\
(log) State IG Rev pp & & & & 0.3695$^{***}$ & & & 0.3498$^{***}$\\
& & & & (0.0388) & & & (0.0380)\\
(log) Fed IG Rev. pp & & & & 0.0049$^{**}$ & & & 0.0033\\
& & & & (0.0023) & & & (0.0021)\\
\midrule
\emph{Fixed-effects}\\
unit & Yes & Yes & Yes & Yes & Yes & Yes & Yes\\
year & Yes & Yes & Yes & Yes & Yes & Yes & Yes\\
\midrule
\emph{Fit statistics}\\
Observations & 12,084 & 12,084 & 11,420 & 11,420 & 13,356 & 12,536 & 12,536\\
R$^2$ & 0.83997 & 0.84139 & 0.85220 & 0.87704 & 0.83621 & 0.84900 & 0.87219\\
Within R$^2$ & 0.17635 & 0.18364 & 0.20940 & 0.34226 & 0.17607 & 0.20933 & 0.33076\\
\midrule \midrule
\multicolumn{8}{l}{\emph{Clustered (unit) standard-errors in parentheses}}\\
\multicolumn{8}{l}{\emph{Signif. Codes: ***: 0.01, **: 0.05, *: 0.1}}\\
\end{tabular}
\par\endgroup
Instrumental Variable Approach
There is a significant endogeneity concern in using total active production and active mines as the treatment variable. Therefore, I have tried two instrumental variable approaches below and aim to add results using production- and employment-based Bartik instruments.
We consider using XX as an instrument affecting education expenditure through property taxes or GDP. We know that property taxes have an endogenous relationship with education expenditure, however, in theory, XX is unlikely to affect education expenditure, except via property taxes. We test this hypothesis below.
As a reminder, the intuition behind the idea is:
Declining vs. Growing Regions
What would be great is to be able to econometrically test when a commuting zone is “declining.” In the first step, it would be good to identify when a commuting zone is declining overall (GDP, poverty, etc) but ideally eventually apply this to the education outcome. My hope is that being able to identify counties that are “declining” we can either use this variable as a covariate or as a central point of analysis. The below analysis looks at state-level variables as a first step (mainly to aid in visual comparison and plotting). Ideally, once a method is decided on this would be applied to commuting zone-level data which would need to be summarise/collated in some way for plotting.
CZ GDP growth conditional on state and national level
[1] 0.2154088
[1] 0.1540881
[1] 0.1194969
Running IV models on declining vs. growing sub-groups
The following implements an employment based Bartik instrument for various industries available from the Quarterly Census of Employment and Wages.
[1] "Downloaded QCEW data for 2004."
[1] "Cleaned temp file."
[1] "Created employment share values."
[1] "Appended national shock variables."
Warning: Removed 2288 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5544 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2992 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 308 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 858 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 528 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 374 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 726 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 770 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5588 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1342 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2068 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1078 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 990 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 748 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 176 rows containing missing values or values outside the scale range (`geom_line()`).
10 Total, all indu.. NAICS 11 Agricultu.. NAICS 21 Mining, q.. NAICS 22 Utilities NAICS 23 Construct.. NAICS 42 Wholesale.. NAICS 51 Information NAICS 52 Finance a.. NAICS 53 Real esta.. NAICS 54 Professio.. NAICS 55 Managemen.. NAICS 56 Administr.. NAICS 61 Education.. NAICS 62 Health ca.. NAICS 71 Arts, ent.. NAICS 72 Accommoda.. NAICS 81 Other ser.. NAICS 92 Public ad..
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp
(log) House Price Index 0.6914*** 0.7542 1.074 -0.4765 0.4016 -0.6195 0.2838 0.4538 4.522 0.5529*** 0.6956 -3.967 1.422 0.3098 0.4402 0.0401 0.2595 -0.0271
(0.1187) (2.582) (0.6927) (0.8539) (0.6486) (0.4157) (0.1909) (0.5339) (7.835) (0.1617) (0.6399) (3.731) (1.279) (0.5756) (0.7036) (0.3581) (0.2751) (0.3720)
(log) Real GDP pc -0.0164 -0.0997 0.3414 0.0790 0.3543** 0.1112 0.0640 -1.041 0.0196 -0.0474 1.228 -0.2409 0.1184 0.0893 0.2067 0.1173 0.1951
(0.6884) (0.1857) (0.2336) (0.1771) (0.1116) (0.0631) (0.1476) (2.156) (0.0523) (0.2252) (0.9761) (0.3741) (0.1510) (0.2100) (0.1055) (0.0799) (0.1056)
Fixed-Effects: -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- --------------------
unit Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
year Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
________________________________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________
S.E.: Clustered by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit
Observations 12,717 11,278 8,751 10,804 12,661 12,426 12,583 12,698 12,638 12,327 8,589 12,071 11,482 12,034 12,320 12,433 12,717 12,717
R2 0.78433 0.77909 0.72206 0.75217 0.81888 0.71278 0.82369 0.81349 -2.2567 0.80722 0.81089 -2.0513 0.58624 0.82553 0.81656 0.82186 0.82389 0.81637
Within R2 -0.12169 -0.20050 -0.59433 -0.29554 0.05177 -0.46957 0.08087 0.02913 -15.915 -0.03038 -0.18121 -14.889 -1.2555 0.08731 0.04593 0.07340 0.08404 0.04494
F-test (1st stage), (log) House Price Index NaN 0.09466 0.98196 1.3275 1.3335 8.4371 43.627 1.7219 0.41478 12.241 1.0043 0.44639 3.1291 7.6945 3.0305 23.495 3.0205 20.341
F-test (1st stage), p-value, (log) House Price Index NaN 0.75834 0.32174 0.24927 0.24821 0.00368 4.13e-11 0.18947 0.51957 0.00047 0.31630 0.50407 0.07693 0.00555 0.08174 1.27e-6 0.08224 6.54e-6
F-test (2nd stage) 4.97e-12 0.04987 1.0694 0.27576 0.19807 2.8965 3.1837 0.31906 7.6684 3.4595 0.50038 6.4333 5.9628 0.67833 0.52728 0.03382 0.18346 0.01348
F-test (2nd stage), p-value 1.0000 0.82330 0.30111 0.59950 0.65629 0.08880 0.07440 0.57218 0.00563 0.06291 0.47935 0.01121 0.01463 0.41018 0.46777 0.85409 0.66843 0.90758
Wu-Hausman -- 0.02834 0.71121 0.53730 0.05196 4.9232 0.30290 0.10225 6.9230 1.3954 0.25809 6.9878 4.3126 0.07581 0.15705 0.47049 0.01078 0.90249
Wu-Hausman, p-value -- 0.86631 0.39907 0.46357 0.81969 0.02652 0.58208 0.74915 0.00852 0.23751 0.61145 0.00822 0.03785 0.78306 0.69189 0.49277 0.91732 0.34213
Wald (IV only) 33.942 0.08532 2.4026 0.31142 0.38329 2.2207 2.2101 0.72250 0.33315 11.695 1.1819 1.1300 1.2356 0.28972 0.39145 0.01253 0.88973 0.00532
Wald (IV only), p-value 5.82e-9 0.77021 0.12117 0.57682 0.53586 0.13620 0.13713 0.39534 0.56382 0.00063 0.27700 0.28779 0.26635 0.59041 0.53155 0.91089 0.34557 0.94188
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
NAICS 54 Professio..
Dependent Var.: (log) Elem.Ed.Exp.pp
(log) House Price Index 0.5529***
(0.1617)
(log) Real GDP pc 0.0196
(0.0523)
Fixed-Effects: --------------------
unit Yes
year Yes
________________________________________ ____________________
S.E.: Clustered by: unit
Observations 12,327
R2 0.80722
Within R2 -0.03038
F-test (1st stage), (log) House Price Index 12.241
F-test (1st stage), p-value, (log) House Price Index 0.00047
F-test (2nd stage) 3.4595
F-test (2nd stage), p-value 0.06291
Wu-Hausman 1.3954
Wu-Hausman, p-value 0.23751
Wald (IV only) 11.695
Wald (IV only), p-value 0.00063
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
[1] "NAICS 54 Professional, scientific, and technical services" NA
% latex table generated in R 4.5.0 by xtable 1.8-4 package
% Tue Jun 3 10:04:38 2025
\begin{table}[ht]
\centering
\begin{tabular}{l}
\hline
industry\_title \\
\hline
10 Total, all industries \\
NAICS 11 Agriculture, forestry, fishing and hunting \\
NAICS 21 Mining, quarrying, and oil and gas extraction \\
NAICS 22 Utilities \\
NAICS 23 Construction \\
NAICS 42 Wholesale trade \\
NAICS 51 Information \\
NAICS 52 Finance and insurance \\
NAICS 53 Real estate and rental and leasing \\
NAICS 54 Professional, scientific, and technical services \\
NAICS 55 Management of companies and enterprises \\
NAICS 56 Administrative and support and waste management and remediation services \\
NAICS 61 Educational services \\
NAICS 62 Health care and social assistance \\
NAICS 71 Arts, entertainment, and recreation \\
NAICS 72 Accommodation and food services \\
NAICS 81 Other services (except public administration) \\
NAICS 92 Public administration \\
\hline
\end{tabular}
\end{table}
Running base models on declining vs. growing sub-groups
.1 .2 .3
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp
(log) Annual Avg. Wkly. Wage 0.2546* 0.1022 0.0183
(0.1037) (0.0984) (0.0993)
(log, l1) Annual Avg. Wkly. Wage 0.2197** 0.1966* 0.1754*
(0.0813) (0.0801) (0.0816)
(log, l2) Annual Avg. Wkly. Wage 0.1611 0.1269 0.1897*
(0.0839) (0.0833) (0.0781)
(log) House Price Index 0.0281 -0.0257
(0.0477) (0.0395)
l1_log_hpi 0.0579 0.0565
(0.0390) (0.0309)
l2_log_hpi 0.0632 0.0282
(0.0383) (0.0337)
l3_log_hpi 0.0820* 0.0801*
(0.0326) (0.0329)
l4_log_hpi -0.0341 -0.0156
(0.0337) (0.0317)
l5_log_hpi -0.1047** -0.0813*
(0.0361) (0.0328)
(log) State IG Rev pp 0.3306***
(0.0575)
(log) Fed IG Rev. pp 0.0077*
(0.0030)
Fixed-Effects: -------------------- -------------------- --------------------
unit Yes Yes Yes
year Yes Yes Yes
________________________________ ____________________ ____________________ ____________________
S.E.: Clustered by: unit by: unit by: unit
Observations 5,418 5,214 5,214
R2 0.82524 0.83229 0.86103
Within R2 0.07186 0.09335 0.24874
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
.1 .2 .3
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp
(log) Annual Avg. Wkly. Wage 0.2736*** 0.2033* 0.2075*
(0.0814) (0.0883) (0.0944)
(log, l1) Annual Avg. Wkly. Wage 0.2204*** 0.1400 0.1524*
(0.0599) (0.0737) (0.0720)
(log, l2) Annual Avg. Wkly. Wage 0.2138* 0.1888* 0.1923**
(0.0955) (0.0730) (0.0693)
(log) House Price Index 0.0488 0.0218
(0.0327) (0.0299)
l1_log_hpi 0.0694 0.0520
(0.0357) (0.0352)
l2_log_hpi 0.0548* 0.0252
(0.0275) (0.0260)
l3_log_hpi 0.0014 0.0080
(0.0277) (0.0262)
l4_log_hpi 0.0064 0.0332
(0.0284) (0.0240)
l5_log_hpi -0.0744** -0.0512*
(0.0280) (0.0230)
(log) State IG Rev pp 0.3685***
(0.0366)
(log) Fed IG Rev. pp 0.0036
(0.0030)
Fixed-Effects: -------------------- -------------------- --------------------
unit Yes Yes Yes
year Yes Yes Yes
________________________________ ____________________ ____________________ ____________________
S.E.: Clustered by: unit by: unit by: unit
Observations 7,938 7,322 7,322
R2 0.81346 0.82416 0.85704
Within R2 0.08664 0.09190 0.26172
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
.1 .2 .3
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp
(log) Annual Avg. Wkly. Wage 0.2511* -0.0216 -0.1060
(0.1257) (0.1148) (0.1176)
(log, l1) Annual Avg. Wkly. Wage 0.2186* 0.1857 0.1967
(0.1095) (0.1063) (0.1022)
(log, l2) Annual Avg. Wkly. Wage 0.1205 0.1423 0.1922*
(0.1019) (0.1123) (0.0945)
(log) House Price Index 0.1490** 0.0638
(0.0507) (0.0434)
l1_log_hpi 0.0091 0.0372
(0.0537) (0.0432)
l2_log_hpi 0.0282 -5.94e-5
(0.0460) (0.0418)
l3_log_hpi 0.0855 0.0836
(0.0449) (0.0443)
l4_log_hpi 0.0005 0.0075
(0.0389) (0.0393)
l5_log_hpi -0.1609*** -0.1413***
(0.0434) (0.0406)
(log) State IG Rev pp 0.2686***
(0.0659)
(log) Fed IG Rev. pp 0.0075
(0.0039)
Fixed-Effects: -------------------- -------------------- --------------------
unit Yes Yes Yes
year Yes Yes Yes
________________________________ ____________________ ____________________ ____________________
S.E.: Clustered by: unit by: unit by: unit
Observations 3,339 3,164 3,164
R2 0.78745 0.80094 0.82736
Within R2 0.07134 0.10056 0.21994
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
.1 .2 .3
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp
(log) Annual Avg. Wkly. Wage 0.1987 0.1783 0.1768
(0.1152) (0.1330) (0.1471)
(log, l1) Annual Avg. Wkly. Wage 0.2616** 0.1725 0.2381*
(0.0962) (0.1157) (0.1127)
(log, l2) Annual Avg. Wkly. Wage 0.2471 0.2193 0.2152*
(0.1388) (0.1114) (0.1011)
(log) House Price Index 0.0764 0.0721
(0.0455) (0.0412)
l1_log_hpi 0.0719 0.0594
(0.0512) (0.0524)
l2_log_hpi 0.0406 0.0073
(0.0387) (0.0380)
l3_log_hpi -0.0151 -0.0279
(0.0374) (0.0348)
l4_log_hpi -0.0251 -0.0142
(0.0483) (0.0418)
l5_log_hpi -0.1272** -0.0912*
(0.0459) (0.0379)
(log) State IG Rev pp 0.2994***
(0.0604)
(log) Fed IG Rev. pp 0.0060
(0.0060)
Fixed-Effects: -------------------- -------------------- --------------------
unit Yes Yes Yes
year Yes Yes Yes
________________________________ ____________________ ____________________ ____________________
S.E.: Clustered by: unit by: unit by: unit
Observations 3,339 2,812 2,812
R2 0.74176 0.76891 0.79942
Within R2 0.08470 0.08674 0.20732
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Growing Regions
Warning: Removed 2288 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5544 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2992 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 308 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 858 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 528 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 374 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 726 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 770 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5588 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1342 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2068 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1078 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 990 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 748 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 176 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2288 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5544 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2992 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 308 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 858 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 528 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 374 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 726 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 770 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5588 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1342 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2068 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1078 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 990 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 748 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 176 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2288 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5544 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2992 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 308 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 858 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 528 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 374 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 726 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 770 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5588 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1342 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2068 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1078 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 990 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 748 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 176 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2288 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5544 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2992 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 308 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 858 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 528 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 374 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 726 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 770 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 5588 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1342 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 2068 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 1078 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 990 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 748 rows containing missing values or values outside the scale range (`geom_line()`).
Warning: Removed 176 rows containing missing values or values outside the scale range (`geom_line()`).
10 Total, all indu.. NAICS 11 Agricultu.. NAICS 21 Mining, q.. NAICS 22 Utilities NAICS 23 Construct.. NAICS 42 Wholesale.. NAICS 51 Information NAICS 52 Finance a.. NAICS 53 Real esta.. NAICS 54 Professio.. NAICS 55 Managemen.. NAICS 56 Administr.. NAICS 61 Education.. NAICS 62 Health ca.. NAICS 71 Arts, ent.. NAICS 72 Accommoda.. NAICS 81 Other ser.. NAICS 92 Public ad..
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp
(log) House Price Index 3.587* 4.607 0.6924 -6.855 0.4563 -1.450 0.4671 -0.5227 -1.298 1.336 1.074 1.134 -1.183 -0.3378 2.270 -0.1860 0.1881 -0.0385
(1.432) (24.53) (0.7798) (26.93) (0.6528) (1.221) (1.164) (3.782) (6.496) (0.7428) (0.7374) (2.034) (3.236) (0.9564) (3.246) (0.3462) (0.7167) (0.9776)
(log) Real GDP pc -0.3364 -0.4860 0.1108 1.289 0.1291 0.4254* 0.1303 0.2740 0.3934 -0.0604 -0.1212 0.0292 0.3510 0.2504 -0.1695 0.2595*** 0.1709 0.2047
(0.2451) (3.723) (0.1312) (4.157) (0.1008) (0.1973) (0.1639) (0.5548) (0.9564) (0.1310) (0.2122) (0.3193) (0.5589) (0.1598) (0.5608) (0.0714) (0.1178) (0.1517)
(log) IG Revenue pp -0.0420 0.2589** 1.154 0.3015*** 0.5024** 0.3484* 0.4056 0.4764 0.2140* 0.2971** 0.2378 0.4617 0.3610** 0.2105 0.3971*** 0.3409*** 0.3636**
(2.157) (0.0822) (2.825) (0.0872) (0.1603) (0.1508) (0.3441) (0.6213) (0.0920) (0.0859) (0.2530) (0.3501) (0.1198) (0.3257) (0.0667) (0.0905) (0.1086)
Fixed-Effects: -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- -------------------- --------------------
unit Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
year Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
________________________________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________ ____________________
S.E.: Clustered by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit
Observations 2,927 2,592 1,802 2,329 2,892 2,864 2,830 2,908 2,864 2,745 1,155 2,529 2,378 2,714 2,691 2,741 2,927 2,927
R2 -1.3731 -2.8559 0.74091 -7.3700 0.79097 0.33880 0.79028 0.72586 0.43022 0.53306 0.54895 0.60859 0.47014 0.76420 -0.01216 0.79459 0.80309 0.79866
Within R2 -8.2123 -13.800 -0.00277 -33.287 0.16014 -1.5021 0.17729 -0.06394 -1.2018 -0.95644 -0.64712 -0.51011 -1.2489 0.05809 -2.8929 0.20041 0.23561 0.21840
F-test (1st stage), (log) House Price Index NaN 0.01095 1.0517 0.03019 0.78716 1.2559 1.1172 0.04489 0.08976 1.0731 0.31378 0.18493 0.67975 1.7623 0.51283 15.823 0.46294 1.2070
F-test (1st stage), p-value, (log) House Price Index NaN 0.91668 0.30526 0.86208 0.37503 0.26252 0.29061 0.83223 0.76450 0.30034 0.57548 0.66721 0.40976 0.18445 0.47398 7.14e-5 0.49631 0.27202
F-test (2nd stage) 2.35e-12 0.21136 0.45162 1.3905 0.15599 2.4060 0.23672 0.01110 0.13883 2.0322 0.41941 0.23240 1.0223 0.19307 2.5951 0.50846 0.01497 0.00163
F-test (2nd stage), p-value 1.0000 0.64574 0.50165 0.23844 0.69291 0.12098 0.62662 0.91611 0.70947 0.15411 0.51736 0.62980 0.31209 0.66041 0.10731 0.47587 0.90264 0.96776
Wu-Hausman -- 0.19351 0.29455 1.3782 0.07694 2.7254 0.13952 0.01669 0.16139 1.6411 0.28485 0.18120 1.2430 0.38229 2.2319 1.2694 0.00145 0.02971
Wu-Hausman, p-value -- 0.66005 0.58739 0.24054 0.78151 0.09888 0.70879 0.89722 0.68791 0.20029 0.59365 0.67038 0.26501 0.53644 0.13531 0.25998 0.96963 0.86316
Wald (IV only) 6.2779 0.03526 0.78828 0.06482 0.48857 1.4112 0.16102 0.01910 0.03995 3.2350 2.1216 0.31105 0.13364 0.12473 0.48916 0.28870 0.06891 0.00155
Wald (IV only), p-value 0.01228 0.85107 0.37474 0.79905 0.48462 0.23496 0.68825 0.89009 0.84159 0.07219 0.14551 0.57709 0.71472 0.72399 0.48436 0.59110 0.79295 0.96858
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Removing outliers - really high-income commuting zones!
As you can see in the scatterplot below, there is a somewhat non-linear relationship between property taxes and elementary expenditure as property taxes collected rise. This happens largely as a result of very high-income commuting zones. Therefore, I exclude any commuting zone that spends more than 28k per pupil to avoid any distorting effects. This removes 12 counties (~2% of the sample) This could benefit from more robust outlier detection. This outlier exclusion weakens our results (and the validity of our instrument choice) in the production-based IV regression. Worth noting and thinking about!!
PANELVAR
---------------------------------------------------
Fixed Effects OLS Panel VAR estimation
---------------------------------------------------
Transformation: demean
Group variable: unit
Time variable: year
Number of observations = 9214
Number of groups = 485
Obs per group: min = 18
avg = 18.99794
max = 19
=====================================================================================================
demeaned_log_real_Total_Educ_Total_Exp demeaned_log_hpi
-----------------------------------------------------------------------------------------------------
demeaned_lag1_log_real_Total_Educ_Total_Exp 0.7384 *** -0.0191 **
(0.0127) (0.0074)
demeaned_lag1_log_hpi 0.1572 *** 1.2273 ***
(0.0223) (0.0130)
demeaned_lag2_log_real_Total_Educ_Total_Exp -0.1526 *** 0.0013
(0.0157) (0.0092)
demeaned_lag2_log_hpi 0.0773 * -0.3512 ***
(0.0343) (0.0200)
demeaned_lag3_log_real_Total_Educ_Total_Exp -0.0256 * 0.0124
(0.0124) (0.0072)
demeaned_lag3_log_hpi -0.0981 *** -0.1435 ***
(0.0206) (0.0120)
demeaned_log_real_gdp_total 0.0823 *** 0.1534 ***
(0.0094) (0.0055)
=====================================================================================================
*** p < 0.001; ** p < 0.01; * p < 0.05
---------------------------------------------------
Fixed Effects OLS Panel VAR estimation
---------------------------------------------------
Transformation: demean
Group variable: unit
Time variable: year
Number of observations = 9214
Number of groups = 485
Obs per group: min = 18
avg = 18.99794
max = 19
===========================================================================================================
demeaned_log_real_Total_Educ_Total_Exp_pp demeaned_log_hpi
-----------------------------------------------------------------------------------------------------------
demeaned_lag1_log_real_Total_Educ_Total_Exp_pp 0.6801 *** -0.0333 ***
(0.0127) (0.0068)
demeaned_lag1_log_hpi 0.1188 *** 1.2537 ***
(0.0243) (0.0131)
demeaned_lag2_log_real_Total_Educ_Total_Exp_pp -0.1088 *** 0.0081
(0.0154) (0.0083)
demeaned_lag2_log_hpi 0.0974 ** -0.3785 ***
(0.0376) (0.0203)
demeaned_lag3_log_real_Total_Educ_Total_Exp_pp -0.0629 *** 0.0138 *
(0.0128) (0.0069)
demeaned_lag3_log_hpi -0.1088 *** -0.1222 ***
(0.0224) (0.0121)
demeaned_log_real_gdp_total_pc 0.0507 *** 0.1416 ***
(0.0116) (0.0063)
===========================================================================================================
*** p < 0.001; ** p < 0.01; * p < 0.05
Property Prices
model 1 model 2 model 3 model 4 model 5 model 6
Dependent Var.: (log) House Price Index gr_hpi log_real_Elem_Educ_Total_Exp diff_log_real_Elem_Educ_Total_Exp (log) Elem.Ed.Exp.pp diff_log_real_Elem_Educ_Total_Exp_pp
(log) Annual Avg. Wkly. Wage 0.5110*** 0.1302 0.2462***
(0.0662) (0.0727) (0.0643)
(log, l1) Annual Avg. Wkly. Wage 0.2052*** 0.1796** 0.1619**
(0.0376) (0.0550) (0.0580)
(log, l2) Annual Avg. Wkly. Wage 0.2789** 0.1149 -0.0153
(0.0885) (0.0759) (0.0600)
(log) Real GDP 0.1368*** 0.0305
(0.0308) (0.0251)
gr_weighted_annual_avg_wkly_wage 0.3141*** 0.0541 0.0341
(0.0332) (0.0505) (0.0547)
l1_gr_weighted_annual_avg_wkly_wage 0.3308*** 0.1774*** 0.1600**
(0.0319) (0.0480) (0.0498)
l2_gr_weighted_annual_avg_wkly_wage 0.2514*** 0.2696*** 0.2124***
(0.0253) (0.0491) (0.0495)
l1_log_real_gdp_total 0.0612***
(0.0171)
l2_log_real_gdp_total 0.1589***
(0.0297)
diff_log_real_gdp_total 0.0140
(0.0166)
l1_diff_log_real_gdp_total 0.0492**
(0.0157)
l2_diff_log_real_gdp_total 0.0030*
(0.0012)
(log) Real GDP pc -0.0164
(0.0243)
l1_log_real_gdp_total_pc 0.0383*
(0.0174)
l2_log_real_gdp_total_pc 0.1031***
(0.0259)
diff_log_real_gdp_total_pc 0.0055
(0.0192)
l1_diff_log_real_gdp_total_pc 0.0276
(0.0169)
l2_diff_log_real_gdp_total_pc 0.0183**
(0.0056)
Fixed-Effects: ----------------------- ---------- ---------------------------- ------------------------------ -------------------- ------------------------------
unit Yes Yes Yes Yes Yes Yes
year Yes Yes Yes Yes Yes Yes
___________________________________ _______________________ __________ ____________________________ ______________________________ ____________________ ______________________________
S.E.: Clustered by: unit by: unit by: unit by: unit by: unit by: unit
Observations 12,612 12,585 11,856 11,855 11,856 11,855
R2 0.96755 0.41767 0.99644 0.09453 0.82837 0.07082
Within R2 0.30497 0.05311 0.17113 0.01198 0.07972 0.00675
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
OLS estimation, Dep. Var.: log_real_Elem_Educ_Total_Exp_pp
Observations: 11,521
Fixed-effects: unit: 609, year: 19
Standard-errors: Clustered (unit)
Estimate Std. Error t value Pr(>|t|)
log_weighted_annual_avg_wkly_wage 0.170435 0.061034 2.79245 5.3957e-03 **
l2_log_real_gdp_total_pc 0.072934 0.030057 2.42654 1.5533e-02 *
log_real_Property_Tax_pp 0.193228 0.015629 12.36361 < 2.2e-16 ***
log_hpi 0.164181 0.019771 8.30411 6.5402e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
RMSE: 0.084828 Adj. R2: 0.83949
Within R2: 0.190245
OLS estimation, Dep. Var.: log_real_Elem_Educ_Total_Exp_pp
Observations: 11,521
Fixed-effects: unit: 609, year: 19
Standard-errors: Clustered (unit)
Estimate Std. Error t value Pr(>|t|)
share_own_discrete::1 = high:log_weighted_annual_avg_wkly_wage 0.158949 0.066604 2.38646 1.7316e-02 *
share_own_discrete::2 = medium:log_weighted_annual_avg_wkly_wage 0.180311 0.061474 2.93311 3.4823e-03 **
share_own_discrete::3 = low:log_weighted_annual_avg_wkly_wage 0.137622 0.066246 2.07746 3.8178e-02 *
l2_log_real_gdp_total_pc 0.071337 0.030426 2.34462 1.9367e-02 *
log_real_Property_Tax_pp 0.195205 0.015886 12.28792 < 2.2e-16 ***
log_hpi 0.164477 0.020041 8.20720 1.3567e-15 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
RMSE: 0.084788 Adj. R2: 0.839612
Within R2: 0.191014
Descriptive Regression Results
In the following set of results, I report descriptive regressions to establish relationships between property taxes, education expenditure, GDP (total, private industry, O&G&mining), etc.
All regression models that follow include TWFE (CZ- and year- fixed effects) and standard errors clustered by commuting zone.
All functional forms in the feols() functions below are of the form “Y ~ X” In the cases in which multiple estimations are included via sw(Xa, Xb, Xc + Xd), the function will return results for “Y~Xa”, “Y~Xb”, “Y ~ Xc + Xd”.
Property Tax ~ GDP
GDP has a highly relevant relationship to property taxes. A 1% increase in GDP (per capita) leads to a 0.38% (0.32%) increase in property taxes collected (per capita).
.1 .2 .1.1 .2.1
Dependent Var.: log_real_Property_Tax log_real_Property_Tax (log) Prop Taxpp (log) Prop Taxpp
(log) Real GDP 0.3854*** 0.1226***
(0.0480) (0.0325)
l(log_real_gdp_total,1) 0.1193***
(0.0274)
l(log_real_gdp_total,2) 0.0697*
(0.0285)
l(log_real_gdp_total,3) 0.0790***
(0.0183)
l(log_real_gdp_total,4) 0.1198**
(0.0384)
(log) Real GDP pc 0.3151*** 0.1212***
(0.0616) (0.0366)
l(log_real_gdp_total_pc,1) 0.0929***
(0.0271)
l(log_real_gdp_total_pc,2) 0.0677*
(0.0328)
l(log_real_gdp_total_pc,3) 0.0731**
(0.0229)
l(log_real_gdp_total_pc,4) 0.0624
(0.0351)
Fixed-Effects: --------------------- --------------------- ---------------- ----------------
unit Yes Yes Yes Yes
year Yes Yes Yes Yes
___________________________ _____________________ _____________________ ________________ ________________
S.E.: Clustered by: unit by: unit by: unit by: unit
Observations 13,356 10,812 13,356 10,812
R2 0.99175 0.99329 0.93467 0.94256
Within R2 0.10787 0.15702 0.06308 0.08956
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Education Expenditure ~ Revenue Sources
The below regressions are included to establish the relationship between education expenditure and its component parts. These regressions simply corroborate what is displayed in the section on Key Relationships in LINK (ie. that the largest form of IG revenue is state funding and “Own Source” revenue is largely sourced from Property Taxes).
.1 .2 .3 .4 .1.1 .2.1 .3.1 .4.1
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp log_real_Elem_Educ_Total_Exp log_real_Elem_Educ_Total_Exp log_real_Elem_Educ_Total_Exp log_real_Elem_Educ_Total_Exp
(log) Rev. Own Sources pp 0.3604***
(0.0190)
(log) IG Revenue pp 0.4469*** 0.4532***
(0.0244) (0.0265)
(log) Prop Taxpp 0.2266*** 0.2871*** 0.2897***
(0.0180) (0.0185) (0.0181)
(log) Fed IG Rev. pp 0.0019
(0.0019)
(log) State IG Rev pp 0.4307***
(0.0283)
log_real_Property_Tax 0.2565*** 0.3014*** 0.3070***
(0.0195) (0.0194) (0.0192)
log_real_Total_IG_Revenue 0.5020*** 0.4853***
(0.0252) (0.0234)
log_real_Total_Fed_IG_Revenue 0.0005
(0.0007)
log_real_Total_State_IG_Revenue 0.4823***
(0.0269)
log_real_Total_Rev_Own_Sources 0.3760***
(0.0191)
Fixed-Effects: -------------------- -------------------- -------------------- -------------------- ---------------------------- ---------------------------- ---------------------------- ----------------------------
unit Yes Yes Yes Yes Yes Yes Yes Yes
year Yes Yes Yes Yes Yes Yes Yes Yes
_______________________________ ____________________ ____________________ ____________________ ____________________ ____________________________ ____________________________ ____________________________ ____________________________
S.E.: Clustered by: unit by: unit by: unit by: unit by: unit by: unit by: unit by: unit
Observations 13,356 13,356 13,356 13,356 13,356 13,356 13,356 13,356
R2 0.89075 0.82859 0.88016 0.87791 0.99566 0.99738 0.99732 0.99763
Within R2 0.45044 0.13778 0.39717 0.38586 0.14427 0.48315 0.47095 0.53223
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Education Expenditure ~ GDP
A 1% increase in GDP pc is associated with a 0.19% increase in education expenditure per pupil, dominated by the effect of GDP from private industry (0.16%). I include here also the GDP generated from the oil, gas, mining, and quarrying sector. The effect is small and statistically insignificant.
.1 .2 .3
Dependent Var.: (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp (log) Elem.Ed.Exp.pp
(log) Real GDP pc 0.1926***
(0.0210)
(log) Real GDP Priv. Industry pc 0.1674***
(0.0182)
log_real_gdp_o_g_mining_quarr_21_pc 0.0155***
(0.0032)
Fixed-Effects: -------------------- -------------------- --------------------
unit Yes Yes Yes
year Yes Yes Yes
___________________________________ ____________________ ____________________ ____________________
S.E.: Clustered by: unit by: unit by: unit
Observations 13,356 13,356 13,356
R2 0.81378 0.81283 0.80330
Within R2 0.06328 0.05847 0.01055
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1